Proc. Nati. Acad. Sci. USA Vol. 89, pp. 5557-5561, June 1992 Immunology

CD4 and CD8 regulate interleukin 2 responses of T cells (Lyt-2/L3T4/p56I&/protein-tyrosine kinase) KAZUHISA TAKAHASHI*t, MOTOMI NAKATAt, TOSHIYUKI TANAKAt, HIROYASU ADACHIt, HIROMITSu NAKAUCHI§, HIDEO YAGITAt, AND Ko OKUMURAt¶ Departments of *Respiratory Medicine and tImmunology, Juntendo University School of Medicine, 2-1-1 Hongo, Bunkyo-ku, Tokyo 113, Japan; tDepartment of Immunology, The Tokyo Metropolitan Institute of Medical Science, 3-18-22 Honkomagome, Bunkyo-ku, Tokyo 113, Japan; and 9Laboratory of Molecular Regulation of Aging, Frontier Research Program, The Institute of Physical and Chemical Research (RIKEN), 3-1-1 Koyadai, Yatabe, Tsukuba, Ibaraki 305, Japan

Communicated by Leonard A. Herzenberg, March 26, 1992

To characterize the T-cell surface molecules ABSTRACT involved in regulation of T-cell interleukin 2 (IL-2) responses, we established several monoclonal antibodies (mAbs) that inhibit IL-2 responses of freshly isolated CD8+ T cells and the IL-2-dependent cell line CTLL-2. Here we show that two inhibitory mAbs are directed against Lyt-2 (CD8a). In fact, all anti-Lyt-2 mAbs tested were able to inhibit the IL-2 response of the Lyt-2- and L3T4-dericient cell line HT-2 after transfection with a Lyt-2 cDNA clone. Similarly, anti-L3T4 mAbs inhibited the IL-2 response of CD4-transfected HT-2 cells. These inhibitory effects of anti-CD4 and anti-CD8 mAbs occur on normal T lymphocytes, since they also were observed with CD4+ and CD8+ T-cell blasts, and are specific for IL-2 responses, since IL-4 responses of CD4- and CD8-transfected HT-2 cells were not affected by the anti-CD4 and anti-CD8 mAbs. The inhibitory effects of anti-CD4 or anti-CD8 mAbs could not be explained by interference with IL-2 binding and depended on CD4 and CD8 crosslinking, because F(ab')2 or Fab plus crosslinking second antibody, but not Fab alone, were effective. A mutant Lyt-2 molecule lacking the cytoplasmic region that mediates p56kk binding could not mediate the inhibitory effect upon crosslinking. These results suggest that CD4 and CD8 mediate negative regulation of T-cell IL-2 responses via cytoplasmically associated p56kk.

Interleukin 2 (IL-2) plays a central role in regulating immune responses (1). IL-2-induced activation and differentiation signals are mediated by specific receptors on the cell surface. The IL-2 receptor (IL-2R) is composed of two polypeptides, p55 IL-2Ra and p75 IL-2RB. p55 or p75 alone constitute lowor intermediate-affinity receptor, respectively, but only the latter is functional in mediating signals. The combination of both subunits constitutes the functional high-affinity IL-2R (2-6). cDNA transfection studies revealed that p55 and p75 were sufficient for reconstituting functional high-affinity IL2Rs on hematopoietic cell lines, but not on nonhematopoietic cell lines (7-13). Similarly, p75 alone reconstituted the functional intermediate-affinity receptor only on hematopoietic cell lines (7, 12). These results suggest that other hematopoietic components in addition to p55 and p75 are required for constitution of the functional IL-2R. In this respect, candidates for IL-2R-associated molecules have been reported (14-17). To further investigate the molecules involved in regulating IL-2 responses, we established several monoclonal antibodies (mAbs) that inhibit T-cell IL-2 responses, two of which reacted with Lyt-2. Here we describe the involvement of CD4, CD8, and their association with p56lck in regulating IL-2 responses of T cells.

MATERIALS AND METHODS Cell Lines. Murine IL-2-dependent T-cell lines CTLL-2 (18) and HT-2 (19) (kindly provided by H. Karasuyama, Tokyo University) were maintained in RPMI 1640 (Nissui Pharmaceutical, Tokyo) supplemented with 10% heatinactivated fetal bovine serum (Bocknek Laboratories, Rexdale, ON, Canada), penicillin (100 units/ml), streptomycin sulfate (50 ,g/ml), and 2 mM L-glutamine (complete medium) containing recombinant human IL-2 (100 units/ml; Shionogi Pharmaceutical, Osaka). P815 murine mastocytoma cells (American Type Culture Collection) were also maintained in complete medium. Antibodies. PC61 (anti-murine IL-2Ra) (20), 53-6.72 (antimurine CD8a or anti-Lyt-2) (21), GK1.5 (anti-murine CD4 or anti-L3T4) (22), and 145-2C11 (anti-CD3) (23) mAbs were provided by H. R. MacDonald (Ludwig Institute), L. A. Herzenberg (Stanford University), F. W. Fitch (Chicago University), and J. A. Bluestone (Chicago University), respectively. Anti-Lyt-2.2 and anti-Lyt-2.1 (allotypic antiLyt-2 mAbs) were gifts from Meiji Milk Product (Tokyo). Anti-Lyt-3.2 and anti-Lyt-3.1 (HO-3.1) (allotypic anti-Lyt-3 mAbs) were kindly provided by E. Nakayama (Nagasaki University). F(ab')2 and Fab fragments of the mAbs were prepared as described (24). Immunizations and Production of mAbs. Adult SD rats (Sankyo Labo Service, Tokyo) were primed with an intradermal injection of 107 T-cell blasts [prepared by culturing BALB/c spleen cells with Con A (5 ,ug/ml) for 2 days] in complete Freund's adjuvant, followed by a booster injection, 10 days later, containing the same number of T-cell blasts in incomplete Freund's adjuvant, and i.p. injection of T-cell blasts in phosphate-buffered saline three times per week. After 10 days, antisera were collected and antibody activities against T-cell blasts were estimated by immunofluorescence. Animals developing a high titer of antibody were given a final i.p. injection of T-cell blasts (2 x 107). Immune splenocytes were prepared 3 days later and fused with P3X63Ag8-653 murine myeloma cells (25). We screened for antibodies that inhibited IL-2 induction of anti-CD3-redirected cytotoxic activity in mouse spleen T cells as described below. mAbs 1C7-3, 2F4, 5D7-1, 5D9-2, and 5H10-1 were selected for their strong inhibitory effects and cloned by limiting dilution. Cytotoxicity Assay. Spleen cells from BALB/c mice were cultured with recombinant human IL-2 (1000 units/ml) in the presence of the indicated mAb (10 ,ug/ml) with or without PC61 (10 ,ug/ml). After 2 days, the cells were harvested and their cytotoxic activity against Fc,, receptor-bearing P815 cells in the presence of anti-CD3 mAb (0.1 ,ug/ml) was tested in a standard 4-hr 5lCr-release assay. The effector/target cell

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: IL, interleukin; IL-2R, IL-2 receptor; mAb, clonal antibody. 1To whom reprint requests should be addressed. 5557

mono-

5558

Immunology: Takahashi et al.

ratio was 50:1, and percent cytotoxicity was calculated as described (26). Proliferative Response of IL-2- and IL-4-Dependent Cell Lines. IL-2/IL-4-dependent cell lines were first deprived of IL-2 for >2 hr in complete medium. They were cultured at 370C for 24 hr in complete medium containing serially diluted IL-2 or IL-4 in the presence or absence of the mAbs (10 ,zg/ml) with or without PC61 (10 ,ug/ml). Proliferation was measured by incorporation of [3H]thymidine (1 ,tCi per well; 1 Ci = 37 GBq) during the last 4 hr of the culture. Proliferative responses of CD81 or CD4+ T-cell blasts, which were induced from CD8+ and CD4+ spleen cells by culturing with phorbol 12-myristate 13-acetate (10 ng/ml), ionomycin (500 ng/ml), and recombinant human IL-2 (50 units/ml) for 5 days, were tested in a similar way. Immunoprecipitation. Surface radioiodination of BALB/c thymocytes and immunoprecipitation were performed as described (27). CD4 and CD8 Expression Vectors and Transfection. The expression constructs containing of the full-length wild-type Lyt-2 cDNA or the mutant cDNA whose cytoplasmic region was truncated (Lyt-2-107A4) were prepared as described (28). The expression construct containing the full-length L3T4 (pSFSVn-LC7) (29) was kindly provided by J. Parnes (Stanford University). HT-2 cells were transfected with those constructs by electroporation (30). Selection with G418 (1 mg/ml) was started 48 hr after transfection. G418-resistant cells expressing a high level of Lyt-2 or L3T4 on the surface were isolated by fluorescence-activated cell sorting. IL-2 Binding Assay. IL-2 binding and Scatchard analysis were done as described (11). Immunoprecipitation and Immunoblotting with Anti-p56ck Antibody. These procedures were performed as described (28).

RESULTS Establishment and Characterization of mAbs That Inhibit IL-2 Responses of Murine T Cells. We first tried to establish a mAb inhibiting IL-2 response of an IL-2-dependent cell line, HT-2, but the resulting mAbs were found to be directed to p55 IL-2Ra, probably due to its extremely high expression on HT-2. Based on our previous observation in human systems that cytotoxic activation of peripheral blood CD8+ T cells is mediated solely by p75 IL-2Rj (31), we next screened for mAbs inhibiting the induction by IL-2 of cytotoxic activity of mouse splenic T cells in an anti-CD3-redirected cytotoxicity assay. We selected five mAbs (1C7-3, 2F4, 5D7-1, 5D9-2, 5H10-1) that did not inhibit the anti-CD3redirected cytotoxicity significantly; control anti-Thy-1 mAb exhibited no inhibitory effect, excluding the possibility that the mAbs simply compete with anti-CD3 mAb for Fc receptor binding to the target cells (data not shown). In addition, these mAbs alone could partially inhibit the IL-2 response of the IL-2-dependent cell line CTLL-2, as well as cytotoxic induction of splenic T cells, and acted synergistically with an anti-p55 IL-2Ra mAb (PC61) in both systems (data not shown). Among these five mAbs, two (5D9-2 and 5H10-1) reacted weakly with CTLL-2 and very strongly with the majority of thymocytes and a subset of spleen cells as estimated by indirect immunofluorescence and flow cytometry. The antigen molecules on thymocytes recognized by these two mAbs were two disulfide-linked heterodimers composed of 38- and 30-kDa subunits or 35- and 30-kDa subunits, as estimated by immunoprecipitation and SDS/ PAGE (data not shown). The distribution and biochemical characteristics of the antigen recognized by 5D9-2 and 5H10-1 were reminiscent of murine CD8. Therefore, we compared the apparent molecular mass of the antigen immunoprecipitated by 5D9-2 and 5H10-1 with that of the CD8

Proc. Natl. Acad Sci. USA 89 (1992)

molecule immunoprecipitated by an anti-Lyt-2 mAb, 53-6.72. These appeared to be identical. In sequential immunoprecipitation studies, preclearing of CD8 from a thymocyte lysate with the anti-Lyt-2 mAb completely depleted the antigen reactive with 5D9-2 and 5H10-1, and vice versa (data not shown). In addition, 5D9-2 and 5H10-1 competed with fluorescein-labeled 53-6.72 for binding to thymocytes as estimated by flow cytometry. Based on these data, we concluded that 5D9-2 and 5H10-1, which inhibited IL-2 responses, are directed to the Lyt-2 (CD8a) molecule. To determine whether the inhibitory effect on IL-2 response is unique for 5H9-2 and 5H10-1, or a general property of anti-CD8 mAbs, we used a panel of anti-Lyt-2 and anti-Lyt-3 (CD8f3) mAbs. All antiLyt-2 mAbs tested significantly inhibited the IL-2 response of CTLL-2, while this effect was not observed with irrelevant allotypic anti-Lyt-2.1 and anti-Lyt-3.1 mAbs (Fig. 1). An anti-Lyt-3.2 mAb did react with CTLL-2 to a similar extent as an anti-Lyt-2.2 mAb, as estimated by flow cytometry, but it did not inhibit the IL-2 response. These results indicate that the inhibitory effect on the IL-2 response is a general property of anti-Lyt-2 mAbs but not of anti-Lyt-3 mAbs. Inhibitory Effects of Anti-Lyt-2 and Anti-L3T4 mAbs on IL-2 and IL-4 Responses of CD4- or CDga-Transfected HT-2 Cells. To further clarify the inhibitory effect of anti-Lyt-2 mAb on IL-2 responses, we introduced Lyt-2 into an IL-2dependent cell line, HT-2, by cDNA transfection. HT-2 originally did not express CD8 or CD4 as determined by flow cytometry and Northern blot analysis (data not shown). HT-2 transfectants expressing high levels of Lyt-2 were isolated by fluorescence-activated cell sorting, and the effect of antiLyt-2 mAbs on the IL-2 response was examined (Fig. 2). IL-2 response of the CD8a (Lyt-2) transfectant, but not that of the original HT-2 cells, was inhibited by 53-6.72 in a dosedependent manner. Similar effects were observed with 5D9-2 and 5H10-1 (data not shown). We also established a CD4 (L3T4)-transfected HT-2 cell line, and observed that its IL-2 response was inhibited to a higher extent by an anti-CD4 mAb (GK1.5) (Fig. 2), indicating that CD4 is also involved in W11,111, fJ-2

r _

- ;.-`. 2

P-ser

VillIllI"a0PA'villaP.,"m EVANNIFFAIVINPA'0VlolPA,VA0.

FIG. 1. Inhibitory effects of anti-CD8a (Lyt-2) and anti-CD83 (Lyt-3) mAbs on IL-2-induced proliferation of CTLL-2 cells. 5D9-2, 5H10-1, and 53-6.72 are rat mAbs against Lyt-2. Lyt-2.2 and Lyt-3.2 are relevant allotypic epitopes expressed on CTLL-2 cells. Lyt-2.1 and Lyt-3.1 are irrelevant allotypic epitopes not expressed on CTLL-2 cells. CTLL-2 cells (104) in 200 jl of complete medium were cultured for 24 hr with recombinant human IL-2 (100 units/ml) in the presence of the indicated mAbs (each at 10 ,tg/ml) with (filled bars) or without (hatched bars) PC61 anti-IL-2Ra mAb (10 ug/ml). Cell proliferation was measured by [3H]thymidine incorporation during the last 4 hr of the culture. Data represent mean + SD of triplicate samples. Dotted line indicates the [3H]thymidine incorporation in the absence of IL-2.

Immunology: Takahashi et al.

5559

Proc. NatL. Acad. Sci. USA 89 (1992)

8, x

~~~~~~~~4

Z 5-

2

0

x

0~~~~~~~~ (-) Intact F(ab')2 Fab Fab+ag

E0. 0

(-) Intact F(ab')2 Fab Fab+ag

Anti-L3T4

Anti-Lyt-2

In

FIG. 3. Requirement for crosslinking of Lyt-2 and L3T4 by anti-Lyt-2 and anti-L3T4 mAbs to inhibit IL-2 response. HT-2 Lyt-2 and HT-2 L3T4 transfectants were cultured with recombinant IL-2 (100 units/ml) in the presence or absence of anti-Lyt-2 (53-6.72) or anti-L3T4 (GK1.5) mAb, respectively. Each antibody or antibody fragment [F(ab')2 or Fab] was added (10 ,Lg/ml). The monovalent Fab fragments were tested with and without anti-rat Ig (alg). IL-2 response was measured as in Fig. 1. Data represent mean + SD of triplicate samples.

HT-2 L3T4 20 10 w v 0 M61r--

w

t%

0.01

0.1

1

10

100

1 10 0.1 IL-4, units/ml

0.01

1000

IL-2, units/ml

100

FIG. 2. Effects of anti-Lyt-2 and anti-L3T4 mAbs on IL-2 and IL-4 responses of HT-2, Lyt-2-transfected HT-2 (HT-2 Lyt-2), and L3T4-transfected HT-2 (HT-2 L3T4). The indicated cells (104) in complete medium were cultured for 24 hr with serially diluted recombinant IL-2 (Left) or IL-4 (Right) in the absence ofantibody (o) or in the presence (10 ,ug/ml) of anti-Lyt-2 (53-6.72) (o) or anti-L3T4 (GK1.5) (A) mAb. Cell proliferation was measured as in Fig. 1. Data represent mean + SD of triplicate samples.

regulation of IL-2 responses. In contrast, control anti-Thy-1 mAb did not inhibit the IL-2 responses of these transfectants (data not shown). In addition to IL-2, HT-2 can also respond to IL4. In contrast to the IL-2 response, however, the IL-4 response of the CD8a- or CD4-transfected HT-2 was not affected by anti-Lyt-2 or anti-L3T4 mAbs (Fig. 2), indicating that the inhibitory effects were specific for IL-2 responses. The inhibitory effects of anti-Lyt-2 and anti-L3T4 mAbs on IL-2 response of the CD8a- or CD4-transfected HT-2 cells were dependent on the extent of CD8 or CD4 crosslinking. More extensive crosslinking using anti-rat immunoglobulin augmented the inhibitory effects of anti-Lyt-2 and anti-L3T4 mAbs (data not shown). Divalent F(ab')2 fragments were equally effective as intact mAbs, but monovalent Fab frag-

ments were inhibitory only after crosslinking by anti-rat

immunoglobulin (Fig. 3). To determine whether the inhibitory effects of anti-CD8 and anti-CD4 mAbs on IL-2 responses of IL-2-dependent T-cell clones could be generated on normal T cells, we prepared CD8+ and CD4+ T-cell blasts from spleen cells by isolating each population by fluorescence-activated cell sorting and activated them with phorbol ester and ionomycin. Anti-Lyt-2 and anti-L3T4 mAbs also significantly inhibited the IL-2 response of CD8+ and CD4+ T cell blasts, respectively (Fig. 4). These results suggest that crosslinking of CD4 as well as CD8 leads to a specific down-regulation of IL-2 responses.

Involvement of Lyt-2-Associated p56Ick in Negative Regulation of IL-2 Responses. Anti-Lyt-2 and anti-L3T4 mAbs used above did not affect the binding of IL-2 to the CD8a- or CD4-transfected HT-2 cells as estimated by Scatchard analysis (data not shown). Therefore, inhibitory effects of these mAbs on IL-2 responses were likely to be directed to the signal-transduction pathway following binding of IL-2 to its receptor. IL-2R-mediated signal transduction has been reported to involve tyrosine phosphorylation (32-35), while CD4- and CD8-mediated signals have been supposed to be mediated by the protein-tyrosine kinase p561ck, which associates with the cytoplasmic portion of these molecules (36-

0

x E 0

:I-

0.01

0.1

1

10

100

1000

0.01

0.1

10

100

1000

IL-2, units/ml

FIG. 4. Effect of anti-Lyt-2 and anti-L3T4 mAbs on IL-2 responses of CD8+ or CD4+ T-cell blasts. CD8+ and CD4+ splenic T cells were isolated by fluorescence-activated cell sorting and cultured with phorbol 12-myristate 13-acetate (10 ng/ml), ionomycin (500 ng/ml), 4nd recombinant IL-2 (50 units/ml) for 5 days. The T-cell blasts were then cultured with serially diluted IL-2 in the presence (e) or absence (o) of anti-Lyt-2 (53-6.72) or anti-L3T4 (GK1.5) mAb at 10 ,ug/ml. Cell proliferation was measured as in Fig. 1. Data represent mean ± SD of the triplicate samples.

5560

Immunology: Takahashi et al.

Proc. Natl. Acad Sci. USA 89 (1992)

HT-2 Lyt-2 HT-2 Lyt-2 -107i 4 HT-2 L-3 I-a,

4.D).-

_N.Po _

~

p56' I-

Al"

33

1

2

3

4

5

6

FIG. 5. Expression of p5ock and its interaction with Lyt-2 and L3T4 in HT-2 transfectants. Immunoprecipitates were examined by immunoblotting with anti-p56ock antibody. Lanes 1 and 2, HT-2 Lyt-2; lanes 3 and 4, HT-2 Lyt-2-107A4; lanes 5 and 6, HT-2 L3T4 transfectants. Lanes 1, 3, and 5 represent one-fifth of anti-p56ock immunoprecipitates; lanes 2 and 4 represent anti-Lyt-2 (53-6.72) immunoprecipitates; and lane 6 represents anti-L3T4 (GK1.5) immunoprecipitates.

38). Therefore, it was feasible to test the involvement of p56lck in the inhibitory mechanisms. We introduced the cDNA encoding a mutant Lyt-2 molecule (Lyt-2-107A4) whose cytoplasmic portion was replaced with 8 nonsense amino acids instead of 28 amino acid residues of the wild-type Lyt-2 and, therefore, is unable to bind to p56lck (28). The mutant Lyt-2 molecule was expressed on HT-2 cells at a level similar to that of the transfected wild-type Lyt-2 (data not shown). The wild-type, but not the mutant, Lyt-2 coprecipitated p56Ick as shown by immunoblotting of the anti-Lyt-2 immunoprecipitates with anti-p561ck antibody (Fig. 5). L3T4 coprecipitated more p56Ick than wild-type Lyt-2, as previously reported (37). By densitometry, 30%, 10%1, and